[PATCH][WAS:bcmai, axi] bcma: add Broadcom specific AMBA bus driver

[Rafał Miłecki]

2011/5/7 Arend van Spriel <arend at broadcom.com>:
> On 05/07/2011 06:49 PM, Rafał Miłecki wrote:
>>
>> 2011/5/7 Arend van Spriel<arend at broadcom.com>:
>>>
>>> On 05/07/2011 03:55 PM, Michael Büsch wrote:
>>>>>
>>>>> Arnd: did you have a look at defines at all?
>>>>>
>>>>> Most of the defines have values in range 0x800 → 0x837. Converting
>>>>> this to array means loosing 0x800 u16 entries. We can not use 0x800
>>>>> offset, because there are also some defined between 0x000 and 0x800:
>>>>> #define BCMA_CORE_OOB_ROUTER           0x367   /* Out of band */
>>>>> #define BCMA_CORE_INVALID              0x700
>>>
>>> Please be aware that the core identifier itself is not unique (in the
>>> current list they are). In the scan the BCMA_CORE_OOB_ROUTER will always
>>> show BCMA_MANUF_ARM (did not look up the proper manufacturer define but
>>> you
>>> get the idea, i hope).
>>
>> Unfortunately, I don't. Could you explain this? How core identified
>> can be not unique? Can 0x800 mean ChipCommon but also SuperPCIeX?
>
> Yes, if ChipCommon is Broadcom core and SuperPCIeX is ARM core (or some
> other). The core identifiers are chosen by a chip manufacturer (eg. Broadcom
> ;-) ). They are not unique by itself so that is why the bcma_device_id
> consists of manufacturer, id, rev, and class. Providing a device table with
> ANY_MANUF would be a bad idea.

OK, we use MANUF in identification... so where is the problem? ;)

My testing patch for b43 "subscribes" for Broadcom's cores only:
static const struct bcma_device_id b43_bcma_tbl[] = {
	BCMA_CORE(BCMA_MANUF_BCM, BCMA_CORE_80211, 0x17, BCMA_ANY_CLASS),
	BCMA_CORE(BCMA_MANUF_BCM, BCMA_CORE_80211, 0x18, BCMA_ANY_CLASS),
	BCMA_CORETABLE_END
};

-- 
Rafał